Every week we receive updates from several channels (kdnuggets – linkedin- many many blogs) to stay tuned to the buzz in the industry as well as read journals and publications as part of everyday work at 7Puentes. This section of our own blog that we inaugurate with this post is an attempt to “compress” all this info as if we were to run a SVN decomposition of it in our heads.
While there is -of course- a lot of noise around Deep Learning (DL) we have to be honest in saying that DL it is still not a core part of our own toolset. However, Tensor Flow has replaced Theano for most of our prototypes and PoC and we’re excited to use Deepdist (yes, it’s about Spark) in our next satellite image analytics project that’s about to begin (the flood of images is coming). We hope DL fulfil all the promises is making and to keep customers satisfied with it’s value. We’ve found out that many prospects in the US have trouble hiring experienced data scientists and in some cases they’ve asked us to interview candidates for them. We confess we were tempted to use a list of questions we’ve received some time ago but fortunately we come up doing our traditional interview and helped our partners with their staffing process. Becoming a data scientist is still very important, specially for professionals with traditional data mining or statistics background. However, from our experience, learning data science by doing is far more effective than an e-learning model. It is good to have someone at your side that can unlock you while working with real data.